Principles of operation

The SYNC_TIME parameter defines a floating time-window (its start and end-times) using PI System relative time notation. PI System relative time notation allows flexible specification of times and ranges relative to the present or other references. PI System relative time notation allows specification of times and ranges relative to the present or other references. It is limited to s, m, h, and d (seconds, minutes, hours, and days respectively).

Before every query execution, the interface reads all the events from the Data Archive that fall within the SYNC_TIME interval, including events at the interval start and end times, and keeps them in memory. Subsequently, the interface executes the corresponding SQL query, and in the Data Archive does the following:

• Adds values with new timestamps

• Replaces values for any existing timestamps

• Deletes the events where it finds no matching timestamps

Be aware of the following:

• Activating this mode causes the interface to read events before every execution from both back-ends: Data Archive as well as the RDBMS. Hence, the WHERE clause of the involved queries, as well as the interval set through the SYNC_TIME. If the sync interval is too long you may overwhelm the PI and RDBMS systems.

• The number of configured queries and the frequency of their execution are important parameters as well, especially when the number of involved queries or execution frequency is too high. You must look at the utilization of relevant resources on both back-ends when using this functionality.

• You should configure the UniInt Health point for each scan class and verify if it finished on time. If not, adjust the scan frequency, shorten the SYNC_TIME interval, or limit the number of executed SQL statements.